NumPy Introduction

By the end of this lesson, you will understand what NumPy is, how to create basic arrays, and why it is essential for efficient numerical computing in Python.

What it is

NumPy (Numerical Python) is a fundamental library for scientific computing. It introduces the ndarray object, a powerful N-dimensional array structure that stores data more efficiently than standard Python lists. Unlike lists, which can hold mixed types and store pointers to objects, NumPy arrays contain homogeneous data stored in contiguous memory blocks. This allows for vectorized operations, where mathematical functions are applied to entire arrays at once without explicit loops.

Key related terms include vectorization (applying operations to whole arrays), broadcasting (handling arrays of different shapes during arithmetic), and dtype (data type specification).

Why it matters

• Performance: NumPy operations are implemented in C, making them significantly faster than equivalent pure Python loops.
• Memory Efficiency: Arrays use less memory because they store raw values rather than object references.
• Ecosystem Foundation: Libraries like Pandas, SciPy, and scikit-learn rely on NumPy arrays as their underlying data structure.
• Concise Syntax: Complex mathematical expressions can be written clearly and briefly using vectorized operations.

Syntax or steps

To use NumPy, import it with the conventional alias np. Create an array from a list using np.array(). Perform element-wise operations directly on the array objects.

Example

import numpy as np

# Create two 1D arrays
a = np.array([1, 2, 3])
b = np.array([4, 5, 6])

# Element-wise addition
sum_array = a + b

# Scalar multiplication
scaled_a = a * 10

print("Array a:", a)
print("Sum of a and b:", sum_array)
print("Scaled a:", scaled_a)

Explanation: First, we import the library. We define a and b as NumPy arrays containing integers. The line a + b does not concatenate lists; instead, it adds corresponding elements (1+4, 2+5, 3+6). Similarly, a * 10 multiplies every element in a by 10. This demonstrates vectorization: one operation handles all elements simultaneously.

Common mistakes

• Confusing lists with arrays: Using [1, 2] + [3, 4] results in [1, 2, 3, 4], whereas np.array([1, 2]) + np.array([3, 4]) results in [4, 6]. Always ensure variables are converted to ndarray.
• Mixed data types: If you create an array with mixed types (e.g., strings and numbers), NumPy casts everything to a common type (usually string/object), losing numerical capabilities. Keep arrays homogeneous.
• In-place modification confusion: Some operations return new arrays, while others modify existing ones. Be mindful of whether you need to assign the result back to a variable.
• Ignoring shape mismatches: Operations require compatible shapes. Adding a (3,) array to a (2,) array raises an error unless broadcasting rules apply.

When to use it

Compare NumPy arrays with standard Python lists to determine the right tool.

Use lists for general programming tasks involving diverse objects. Use NumPy whenever you are performing calculations on large sets of numbers.

Practice

Guided Exercise: Create a NumPy array representing the squares of numbers 1 through 5. Hint: Use np.arange(1, 6) to generate the base numbers, then multiply the array by itself (arr * arr).

Challenge: Given two arrays x = np.array([1, 2]) and y = np.array([3, 4]), calculate the dot product manually using NumPy operations without using np.dot(). Expected output: 11.

Quick check

Question: What happens if you try to add a NumPy array of integers to a Python list of integers?

Answer: It works seamlessly. NumPy automatically converts the list to an array internally before performing the element-wise addition.

Summary

NumPy provides the ndarray structure for efficient, homogeneous numerical data storage. Its primary advantage is vectorization, allowing fast mathematical operations on entire datasets without explicit Python loops. Mastering basic array creation and arithmetic is the first step toward advanced data science workflows.